Conventionally, in an authentication technique using biometric information, two different processing of “Authentication” and “Identification” are performed.
The “Authentication” includes registering the biometric information of a user having a specified ID, and obtaining once again the biometric information of the user so as to determine whether the biometric information matches the previously registered biometric information. This is called “one-to-one verification”.
The “Identification” includes registering biometric information of many persons and obtaining once again the biometric information of a target user so as to determine which one of the registered persons the target user matches. This is called “one-to-many identification” These two processing have both an advantage and a disadvantage, respectively.
That is, in “one-to-one verification” in which a user specifies an ID that can identify a particular individual, authentication processing can be performed at high speed and with high accuracy, while an ID specification operation is cumbersome for a person to be authenticated and there is a possibility that the ID would be forgotten.
On the other hand, in “one-to-many identification” in which an ID that can identify a particular individual is not used but which one of a plurality of persons is identified, it is possible to eliminate the need to use the ID, while authentication processing cannot be performed at high speed and with high accuracy.
As to a conventional “one-to-many identification” technique, in order to achieve high speed and high accuracy authentication processing, there has been used a method that narrows down the amount of information to be matched by using auxiliary information such as an ID or password.
For example, there is known a method in which part of input information is used as a hash. This method includes: setting arbitrary bits of individual identification information (ID) as selection information; searching for individual identification information including the selection information corresponding with selection information input through a selection information input means; and matching biometric information corresponding to the searched individual identification information which are stored in an individual information storage means and biometric information input through a biometric information input means so as to identify an individual (refer to, e.g., Patent Document 1).
Further, there is known a technique in which biometric information and password that a user can freely set are combined, and the amount of information to be searched for in the “one-to-many identification” is narrowed down by the password (refer to, e.g., Patent Document 2).
Further, as a technique using auxiliary information, there is known a technique that uses a function of simulating authentication accuracy. In this technique, in the case where the authentication accuracy does not satisfy a predetermined criterion, a user is required to input a password as auxiliary information for verification of identity (refer to, e.g., Patent Document 3).    Patent Document 1: Japanese Laid-Open Patent Publication No. 2000-123179    Patent Document 2: Japanese Laid-Open Patent Publication No. 2000-215172    Patent Document 3: Japanese Laid-Open Patent Publication No. 2000-259278